Artificial Intelligence in Medicine: Deep Learning Supports Active Surveillance
The Declaration of Independence cites the pursuit of happiness as one of the inalienable rights with which we are all born. Pursuing your happiness is easy when your life is going along as usual. However, an unpredictable unpleasant event such as a diagnosis of prostate cancer (PCa) slams the brakes on your normal quality of life. Your routine is now interrupted by doctor appointments, tests, choices and decisions, the possibility of an invasive intervention followed by a recovery period, and potential side effects that demand learning new coping strategies. So much for pursuing happiness!
Don’t despair, there’s hope for minimizing such disruption. A growing number of men diagnosed with early stage, low-risk PCa are finding that less is more. The less intervention, the more normal their life, and the more they can pursue happiness. They minimize interruptions by opting to go on Active Surveillance (AS). This allows them to go about their usual lifestyle while deferring treatment. Of course, AS itself is not without scheduled obligations that can distract from a patient’s happiness enterprise. Periodic blood tests and MRI scans, essential for tracking the cancer’s activity, may be a source of anxiety that temporarily interferes with enjoying high quality of life.
Deep Learning and peace of mind
Easing stress and worry during AS may soon be possible, thanks to a subset of Artificial Intelligence (AI) called Deep Learning (DL). DL algorithms are of increasing interest and use in detecting and diagnosing PCa. Such programs are modeled after the brain’s neural networks, and require less human intervention and Machine Learning algorithms once they are in place. As the online learning center Flatiron School explains, “Complex, multi-layered ‘deep neural networks’ are built to allow data to be passed between nodes (like neurons) in highly connected ways. The result is a non-linear transformation of the data that is increasingly abstract.”
A study presented at the 2022 annual meeting of the American Urological Association (AUA) revealed the potential for DL interpretation of prostate MRI scans to predict the course of PCa in patients on Active Surveillance. The ability to anticipate and predict how stable the cancer will be can ease fear and worry over how the disease is behaving. With the aid of DL, multiparametric MRI (mpMRI) can provide timely information which may prevent a monitoring biopsy, and thus calm anxious minds.
The National Cancer Institute study involved 49 patients on AS. Their monitoring included an average of four multiparametric MRI scans during over roughly five years. Starting at baseline, DL was used through time to track the number of lesions and total disease burden. The objective was to identify factors associated with disease progression in order to develop a predictive DL program. Over time, 24 of the patients experienced disease progression while 25 did not, so the team had a basis for comparison.
As explained in a news report on the paper presented to the AUA by Dr. Michael Daneshvar, a greater percentage of the cases of progression had three characteristic factors: high-risk lesions, a greater burden of disease, and a faster tumor growth rate (0.17 cc/year vs. 0.04 cc/year for those without disease progression).
Why is this helpful for patients on AS? Using DL to swiftly pinpoint these factors can make the difference between staying on AS without needing a monitoring biopsy, or having a biopsy to detect disease progression as early as possible to determine if it’s time to move to definitive treatment. The advantage of adding DL tools to MRI interpretation is the speed and accuracy they offer over human interpretation. The sooner a patient knows he can remain on AS, the sooner he can continue to maintain the quality of life that helps him pursue his happiness without anxiety.
Dr. Daneshvar acknowledged that is a pilot study. The patients in it will continue to be monitored, and larger scale research is warranted. Since our Center is in the vanguard of AI use with prostate MRI, all of us will eagerly be watching for future studies that confirm NCI’s work.
NOTE: This content is solely for purposes of information and does not substitute for diagnostic or medical advice. Talk to your doctor if you are experiencing pelvic pain, or have any other health concerns or questions of a personal medical nature.
- Artificial Intelligence, Prostate imaging